On 10/9/2006 8:48 PM, Michael wrote:
> Hi,
> 
> I?ve been going through the tutorial for Replicating Your First Database 
> and I have a few questions before I touch our  production database.

If you don't have a test environment that's a more or less exact copy of 
your production system, there has never been a better time to build one 
than right now.

> 
> 1)       in the set add table part of the first script, is there a way 
> for me to determine in postgres, which of the tables in a db ( that I 
> didn?t design) that do not have a primary key (sort of like the history 
> table in the example) before I run this script. I don?t want to have to 
> trip over an error each time, I?d like to pre-empt this with a key = serial.

The serial keys added by slony are considered a design flaw to begin 
with. Do not use that feature. Make sure that all your tables have a 
primary key before you start working with slony on that database.

> 2)       in the third script (the one that validates that the replicated 
> db is the same as the master), should I include sequences as well as 
> tables for this comparison?

You can do that. Keep in mind however that gaps in sequences are ok. So 
as long as you find a replica to be ahead (can happen in crash and 
failover scenarios with actual loss of committed transactions), this is 
still in accordance with PostgreSQL's semantics of sequences.


Jan

-- 
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#================================================== JanWieck at Yahoo.com #